Error >> java.io.FileNotFoundException: Cannot access /tmp: No such file or directory

[GwangSeok Song]

Hi, everyone

I have installed Rhipe-0.74.0, with R-3.1.2, Hadoop-1.0.4 and protobuf-2.4.1 in multi node cluster.

When i execute R script, the same message appears below.

("RJavaTools", "Ljava/lang/Object;", "invokeMethod", cl,  :
  java.io.FileNotFoundException: Cannot access /tmp: No such file or directory.


R script is ...

############
library(Rhipe)
rhinit()

rhput("pg2701.txt", "/user/f515/test/")

map <- expression({ words_vector<-unlist(strsplit(unlist(map.values),"")) llapply(words_vector,function(i) rhcollect(i,1)) })

reduce <- expression( pre={total<-0},  reduce={total<-sum(total,unlist(reduce.values))},  post={rhcollect(reduce.key,total)})

Job <- rhwatch(map=map,reduce=reduce,inout=c("text","sequence"),ifolder="/user/f515/test/", ofolder="/user/f515/test/out",jobname="word_count")

rhex(job)
############

/tmp is not used anywhere. and environment variable declarations also have fully.

In the script is also wrong, I tried to run the other examples.


############
  make.words <- function(N, destination, columns = 5, symlength = 5, mulfac = 1) {
    map <- expression({
      Nf <- mulfac*length(map.values)
      replicate(Nf, { ##NF Lines of text
        f <- sapply(1:columns, function(n) paste(sample(letters, symlength), collapse = ""))
        rhcollect(NULL, f)
      })
    })
    z <- rhwatch(map=map, reduce=0, input=N, output=rhfmt(type='text', folders=destination,writeKey=FALSE)
                 ,param=list(mulfac=mulfac,symlength=symlength,columns=columns),read=FALSE)
    z
  }

result <- make.words(c(1000,10),"/user/f515/test/sampletext", columns=10, symlength=7)
############

The above error message appears the same, though.

How can I correct this? Please help.

※ In 1777 permission was given /tmp

Thanks.

[Saptarshi Guha]

by default,
rhoptions()$HADOOP.TMP.FOLDER

is set to /tmp/

set it to something else or make a /tmp folder on the HDFS

--

---
You received this message because you are subscribed to the Google Groups "rhipe" group.
To unsubscribe from this group and stop receiving emails from it, send an email to rhipe+un...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[GwangSeok Song]

Thank you for the answer. To solve this problem, I was worried for 4 hours.

Like your answer, This problem could be resolved by creating a tmp folder on HDFS.

hadoop fs -mkdir /tmp

Once more I thank you for your help.


2015년 4월 4일 토요일 오전 2시 19분 5초 UTC+9, Fishtank 님의 말: